Independently of the component used to introduce a divergence angle between the two probing beams of straightness interferometers, their uncertainty is limited by three main errors linked to each other: their resolution, the influence of refractive index gradients and the topography of the straightness reflector. The larger the divergence angle the higher is the resolving capability, but also the potential influence of the other two error sources. A fully fibre-coupled heterodyne interferometer was successively equipped with three different wedge prisms to investigate the optimal divergence angle under laboratory conditions. For that, the straightness interferometer was qualified with the Nanometer Comparator, which is a one-dimensional line scale interferometer with an additional straightness measurement capability. This feature is based on the traceable multi-sensor method, where an angle measurement embodies the “straightedge”. Therefore, the qualification of the straightness interferometer was also a comparison of two different straightness measurement methods. The influence of the refractive index gradients of air did not increase with interspaces between the probing beams larger than 11.3 mm. Therefore, over a movement range of 220 mm, the lowest uncertainty was realized with the largest divergence angle. The dominant uncertainty contribution arose from the uncorrected mirror topography determined with sub-nanometre uncertainty with the Nanometer Comparator.